News & Events

‘Alarm fatigue’ a factor in 2nd death

‘Alarm fatigue’ a factor in 2nd death

UMass hospital cited for violations

By Liz Kowalczyk

Globe Staff 

The second patient death in four years involving “alarm fatigue’’ at UMass Memorial Medical Center has pushed the hospital to intensify efforts to prevent nurses from tuning out monitor warning alarms.

Nurses exposed to a cacophony of beeps may no longer hear them or begin to ignore them, and that’s what appears to have happened in the latest case: A 60-year-old man died in an intensive care unit after alarms signaling a fast heart rate and potential breathing problems went unanswered for nearly an hour, according to state investigators who reviewed records at the hospital.

The death occurred in August 2010 but was not reported to the state Department of Public Health until this spring. The state cited various violations by the hospital, including not responding to alarms “in a timely manner.’’

The fatality was remarkably similar to another at the Worcester teaching hospital four years ago, when nurses didn’t hear or ignored alarms indicating that the battery on an elderly woman’s monitor needed to be replaced. After the battery failed, 77-year-old Madeline Warner suffered cardiac arrest and the alarm didn’t sound.

Warner’s death in 2007 led the hospital to adopt aggressive measures to improve nurses’ responses and tackle alarm fatigue, which can occur when nurses hear alarms – many of them false – all day long. But the new death shows the problem continues, as it does at hospitals nationwide. It has led to at least 200 patient deaths since 2005 and likely hundreds more, according to a Globe investigation published earlier this year.

‘I don’t think any hospital has fully solved this problem.’

“I don’t think any hospital has fully solved this problem,’’ said Maria Cvach, an assistant director of nursing at Johns Hopkins Hospital in Baltimore and a specialists on alarm fatigue. At Hopkins, which is known for its intense focus on reducing false alarms and alarm fatigue, Cvach said she still sees nurses “not hearing or not answering alarms’’ because “they go off all the time.’’

Executives at UMass Memorial declined to comment about the latest death. The Globe obtained the Department of Public Health report on the case through a freedom of information request. The document does not say who disclosed the death to the state or provide a reason for the reporting delay, but the Globe’s stories on alarm fatigue, which included an account of Warner’s death, appeared several weeks before the health department was called.

The state inspectors’ report, which omits the patient’s name and other identifying details, indicates that the patient also was the apparent victim of a medication error. It does not make clear the extent to which that error or alarm fatigue contributed to the patient’s death.

The patient had been cutting down a tree and was struck in the head with a large branch in July 2010, breaking bones in his face and causing bleeding in his brain. After surgery at UMass Memorial, the patient was often restless and agitated. He kept removing a monitor that measured whether his blood had enough oxygen, triggering an alarm to sound much of one day – which could have been a factor in desensitizing nurses.

As a result, staff tried to calm him with medications. A resident told a nurse to give the patient the antianxiety drug Ativan in small doses up to 5 milligrams. But the resident didn’t write the order that way, and the nurse gave the 5 milligrams all at once.

The report does not say if this medication slip-up was a factor in what occurred 60 minutes later. The patient’s monitors began sounding alarms around 5 p.m., warning that his heart rate and breathing had sped up, and his blood oxygen level had fallen. Eventually, a critical alarm, which sounds more shrill, warned at about 6 p.m. that the patient had stopped breathing.

“There was no evidence that nursing had responded to the alarms during this period of time,’’ state investigators found.

The patient’s nurse told the investigators that the patient was OK when last seen at about 5:15. The nurse was updating a medical record in another patient’s room and thought someone else had responded to the alarms, the nurse later told investigators. After the critical alarm sounded, the nurse went to the man’s room; his face looked gray, and the nurse called emergency personnel, who performed CPR and connected the patient to a ventilator. A CT scan later showed a brain injury due to lack of oxygen and the patient never improved. His family withdrew life support several days later.

Dr. Robert Klugman, UMass Memorial’s chief quality officer, said in an interview that monitoring saves thousands of lives. He said “the systems are expected to be in place where people are aware if someone is tied up, and other people will respond. It’s pretty generally adhered to.’’

The hospital assured the public health department that it is taking steps to improve care, including holding monthly drills in the ICUs timing how long it takes nurses to respond to alarms; and arranging seminars and webinars on reducing false alarms and guarding against alarm fatigue, the report said. Hospital spokesman Robert Brogna said the educational programs are voluntary for the hospital’s 2,100 nurses.

After Warner’s death, the hospital began sending low-battery warnings, as well as alarms for many potential life-threatening changes in a patient’s condition, directly to nurses’ cellphones and pagers.

Other hospitals have found alarm fatigue to be a stubborn problem, despite their best efforts to reduce it.

A study at Hopkins, led by Cvach in 2007, documented an average of 942 serious alarms per day – about one critical alarm every 90 seconds – on one 15-bed unit. Other studies have found that more than 85 percent of alarms are false, meaning the patient is not in any danger.

Hopkins has been trying to eliminate false alarms caused by patient movement and by slight changes in a patient’s condition that don’t require additional care.

It reduced alarms on one unit from 300 per patient per day to 100 per patient per day, Cvach said. But “even if there are 100 alarms per patient per day and a nurse has three patients, that’s still a lot of alarms a nurse has to deal with,’’ she said. “We still haven’t conquered the problem.’’

Like many nurses and doctors, Cvach believes one key solution is better machines, which take into account multiple measures of a patient’s health – heart rate, respiratory rate, blood pressure, and blood oxygen level – to more accurately gauge whether a patient is really in crisis before an alarm sounds.

Theresa Gallivan, associate chief nurse at Massachusetts General Hospital, said monitor makers need to improve their products.

Following an alarm fatigue-related death at Mass. General last year, which was first reported by the Globe, the hospital has overhauled its system for monitoring patients.

“Do we have the perfect system that has cracked the code on this? Absolutely not,’’ Gallivan said. “There is such a mismatch of technology and what we are trying to achieve.’’

Liz Kowalczyk can be reached at